borderline category is more unlikely than likely. If there are exceptions, they are likely to be lesions in vertical growth at thin level IV invasion (with minimal lymphoid infiltrates in the area of level IV invasion) .
Lesions in this borderland might be appropriately classified as: borderline melanocytic neoplasia of indeterminate malignant potential and subdivided into two subcategories:
1. dysplasia category (moderately severe to severe atypia at levels I & II) (nuclear grade III-IV: high grade dysplasia) .
2. melanoma category (at least moderate atypia with markers for host immune response, and with a vertical growth component [level III or deeper]) (nuclear grade II-III, rarely IV) .
Tolerance in the face of conflicting reports: Access to all available data promote the richness and utility of evoked virtual images. Integration of current, and antecedent clinical data provides perspective. In the final analysis, a selection of the most appropriate virtual images is simplified, if the clinical course has been defined during a period of observation. A pathologist who renders a diagnosis of malignant melanoma late in a clinical course should temper his criticisms of any earlier, but less definitive, interpretations by others of sequential histologic material (he should also keep the relativity of the diagnosis of melanoma in the forefront of his deliberations). A flood of historical perspectives is the luxury of retrospection. In its wake, the self-righteousness of experts easily overflows.
A pathologist may be asked to interpret an area of possible regression. If he has no insight as to the nature of the lesion as it existed prior to regression, and no assurances, that the subsequent clinical course will demonstrate regional nodal metastases, his pertinent virtual images will likely be wideranging. On the other hand, another pathologist, who may be asked to interpret sections of such a problematic site, and has information that the patient also has regional, or distant, metastases of melanoma, will have access to virtual images of a different order. The interpretations of the changes in the altered primary site by the two pathologists may differ greatly. From the pattern of real images in the primary site, the original pathologist may have evoked virtual images of a senescent lichenoid dermatitis (e.g., a fixed drug eruption), or perhaps poikiloderma of civatte, Berloque's dermatitis, a senescent lichenoid pigmented keratosis, regression of a halo nevus, regression of premalignant melanocytic dysplasia, or even an area of regression in either a T-cell, or keratinocytic, dysplasia. With knowledge of the clinical course, the second pathologist, who has access to real images in both the site of regression, and the metastases, may evoke only virtual images of melanoma to the exclusion of all other possibilities. In making accommodations for the more extensive data, the second pathologist, during his interpretations, will have imposed, with a feeling of certainty, a limited set of virtual images and, thereby, will have modified, and focused, the significance of the real images. To have the advantage of another pathologist in this manner, and to then testify that the performance of the other pathologist is a deviation from a "standard of care," is an insensitive conceit.
Variants of melanoma (other than the common forms): In common mela- nomas, cytologic variations in vertical growth components are mostly ignored, even in the category of superficial spreading melanoma. On the other hand, cytologic atypia, and abnormal patterns, in which nests and fascicles are abnormally aggregated, have not been manipulated in a fashion which acceptably defines uncommon variants of melanocytic neoplasia, including melanoma variants. Attempts to define such variants, and to characterize them as minimal deviation melanomas have not been enthusiastically received. In part this reflects the primacy of the current classifications of melanoma in which emphasis is either given to radial growth patterns, or the default position of “one melanoma, biologically and histologically is adopted.
Vertical growth: Dimensionalities, as utilized in the manipulation of virtual images, provide the distinctions between melanocytic neoplasias that are dysplastic in character, and those which are melanomatous in character. The presence or absence of a vertical growth component provides defines the distinctions. An atypical melanocytic neoplasm with a vertical growth component is, with rare exceptions, a melanoma, and an atypical melanocytic neoplasm with no such component and only a "two dimensional" dermal component is a dysplasia, regardless of the degree of atypia, the upper migration of neoplastic cells into the epidermis, or the presence of occasional nests of atypical neoplastic cells in a widened, inflamed papillary dermis.
The cytologic variations in dysplasias provide evidence of neoplastic progressions; they are morphologically accessible by an evaluation, and an assignment, of degrees of dysplasia (atypia). The segmentation of the portion of the neoplastic progressions that might be characterized as the melanocytic dysplasia continuum is sensitive to individual biases, but at least 3 designations give recognition to epidermal patterns which might be characterized as a common final pathway. The list includes severe compound dysplasia, "melanoma" in situ, and SSM at level II. In the concept of MDM, these three segments, at a transition zone along the neoplastic continua, qualify as a common final pathway. The succeeding options - once the pattern of the common final pathway has been expressed - include: 1), access to deeper levels (progressive increase in number of nests and stratification of nests), and 2), alterations in the manner in which cells react with their stroma to promote the survival of closely aggregated nests of neoplastic cells in the dermis (i.e., the patterns lend themselves to the structuring - from a review of the two dimensional real images of a histologic section - of virtual images relevant to a vertical growth component.
Options 1 & 2 define the steps leading to the emergence of a vertical growth component. Both are available to dysplasias which have evolved to the stage of the common final pathway. In the concept of MDM, the options are not confined to neoplasia at the limits of the spectrum of the dysplasias. This pathway to vertical growth can be accessed before the histologic patterns are those of the common final pathway (i.e., the pathway to vertical growth is accessible prior to the emergence of the common final pathway). That some lesions enter vertical growth without evidence of the common final pathway on either side of the dermal-epidermal interface is most readily appreciated in the setting of lentigo maligna melanoma. In this variant, cells in the vertical growth components may show only mild to moderate atypia. If such an option is available in the setting of actinic melanocytic dysplasias, there is no reason to dismiss the likelihood that a dysplasia in some other setting might enter vertical growth from a population of cells showing a nuclear grade less than that of the common final pathway. These options are embodied in the concept of MDM.
Virtual images: on recall, our stores of images, all structured to relate to histologic patterns, are virtual images. To have utility, virtual images must be translated into, and used as, language. They are the tools of cerebration. The real images of microscopy which are inherent in the structure of a histologic section are integrated, and modified by interaction with the virtual images of the microscopist.
We may not be comfortable with the proposal that real images are significantly modified during histologic examinations but, if a microscopist is alert to inherent hazards, he will surely have experienced the phenomena whereby he will have quickly evoked a diagnosis, and confirming histologic features, only to later modify not only the interpretation, but even the related "real" images, as they were first perceived. He will likely be appalled at his initial perambulations. Often, we see what we want to see, or what we have been programmed (by attention to clinical details, laboratory reports, etc.) to see, or even what we have been told we should see.
If we have been programmed to see "melanoma" in situ, then we become comfortable with the diagnosis without seriously questioning the appropriateness of the designation. Similarly, superficial spreading melanoma at level II has become ingrained in our store of virtual images. On the other hand, few of us are comfortable in restricting the diagnosis of melanoma to lesions in which the real images embody the virtual images of vertical growth (a set of virtual images laden with implications in regard to neoplastic progressions, and even prognosis) but this approach is the basis for the concept of MDM.
A portion of a neoplastic spectrum, currently accepted by most observers as within the realm of "melanoma," was first isolated by defining its limits; was then labeled "melanoma" at level II; and finally was analyzed statistically. In the analysis, it has been found that this segment will rarely portend a significant risk for progressive disease. This portion of the common, dysplasia-melanoma spectrum is thin, and two dimensional. The same potential is inherent in portions of the spectrum that is thin, and three dimensional (i.e., portions of the spectrum in which lesions measure less than 1 mm in vertical dimensions at the "thickest" portion of any vertical growth component), with only a minimal increase in risk to the patient. Since a prediction of progressive disease for these two portion of the spectrum is most likely to be inaccurate (whereas, in predicting 5 year survivals, the degree of accuracy tends to take on the qualities of prescience), it would be advantageous to substitute a designation which conveys our limitations for these two segments of a neoplastic spectrum.
The utility of real and virtual images is found in the degree of concordance between the predicted, and the actual course of problematic lesions. If the real images of a real melanoma have been supplanted by imposed, false virtual images and if, in the imposition, a malignant neoplasm has been characterized as benign, the subsequent course likely will reveal the masquerade. With the appearance of local recurrence, skin metastases (satellite lesions), regional nodal metastases, or distant metastases, the true nature of a problematic lesion will have been exposed.
In the prosecution of malpractice claims, the testimony of experts may be critical, but in some cases the opinions of experts are sufficiently opposed for one to negate the other. In the exercise of an expert's prerogatives, a critical examination of records, and patterns makes a mockery of what constitutes the daily practice of pathology. The acceptance of an invitation to participate in a malpractice case as an expert is a confession of the newly empowered expert's prejudices. In developing his position, an expert usually touches only briefly on facts, and relevant images; much of his approach will be concerned with the manipulation of virtual images which relate to statistical analyses rather than to the case at hand. To accept the role of "expert" is an admission that the universal human frailty is not among the weaknesses of the respective expert. Hypocrisy is the unacknowledged weakness of all who accept the label, expert.
The following scenarios are excerpts of the positions taken by experts in some medico-legal proceedings:
If a primary melanoma, mistakenly diagnosed as benign nevus, is associated, either synchronously or sequentially, with regional nodal, or blood-borne, metastases, but does not recur locally following the diagnostic biopsy, then, by inference, the lesion in its primary site must have been eradicated by the local excision. Prior to the local excision, such a lesion must have attained the properties of a fully evolved, metastasizing melanoma (whatever its vertical dimensions, and whatever the imposed label). In this scenario, the subsequent appearance, and the site, of the metastases provide a retrospective definition of the clinical stage of disease prior to the initial biopsy of the primary melanoma; by inference from these clues, the problematic primary lesion at the time of its initial excision would be revealed as having been either occult, clinical stage II, III, or IV. The respective patient would have been committed to progressive disease prior to the excision of the primary melanoma. Even the subsequent appearance of only skin metastases (in patterns of satellitosis) would provide insight into the stage of the disease prior to the complete excision of the primary lesion. Occult metastases in the skin at the time of the excision of the primary lesion would come to be documented only by subsequently becoming overt. In their occult stage, they would, however, have preceded the excision of the primary lesion.
Virtual images - factual or fanciful? Factual virtual images are congruous with observer phenomena. Fanciful virtual images are evoked to provide an explanation for clinical and histologic paradoxes. The concept of "in transit" metastasis (fanciful virtual images) (12) has found application in the manipulation of clinical data following a misdiagnosis of a primary melanoma, particularly in those examples in which the primary site was not promptly re-excised following biopsy. Virtual images of malignant cells free in lymphatics, and uninfluenced by the direction of the flow of lymph ("in transit") lend themselves to a manipulation, beyond the intent of the surgeons who developed the concept: - a concept for evaluating the utility of delayed node dissections.
The following scenario is an example:
As an initial procedure, a melanocytic lesion, clinically of problematic nature, would be conservatively, but completely, excised. If the lesion had been truly a melanoma, but its nature had not been appreciated, and the lesion had been interpreted as benign, the pathologist and then the clinician might assume that there would be no indication for additional treatment. If the lesion did not recur locally, but the subsequent clinical course was characterized by regional nodal, or distant, metastases; and if a failure to re-excise the local site is then cited as evidence of negligence, fanciful virtual images of melanoma cells free in the lymphatic circulation, but all clustered just beyond the surgical limits of the primary excision site, would be required as justification for an indictment. In these virtual images, cells in some manner would resist the flow of, and swin against, the stream of lymph to maintain a static position; they would not be implanted. Following the biopsy of the primary lesion, and after an arbitrary interval greater than that required by convention for a re-excision of the biopsy site (had the primary lesion been correctly diagnosed following the histologic examination of the biopsy material), the melanoma cells then would follow the flow of lymph, and colonize the lymph nodes. They would not colonize the site of the primary lesion. With the creation of such a fanciful scenario, an expert might then state that, if the potential had been appreciated by pathologists and oncologists, the fanciful, static, endolymphatic component could have been eradicated by a re-excision of the the biopsy site (even with the limitation that the width of the excised specimen might have been no more than 1 cm). In this setting, and approach, the patient might be characterized as having been harmed firstly by the histologic misdiagnosis and, secondly, by the failure of the clinician to more widely re-excise the primary site. In fact, the nature of the original lesion would have been defined by both historical perspective, and fanciful virtual images. When confronted with this scenario, all retrospections of virtual images, interpretations, and prognostications must accommodate an admission that the patient was committed to progressive, fatal disease even prior to the original biopsy. The clinical course would have been independent of the manner in which the primary lesion was managed. The biologic determinants are to be found in neoplastic cells, and their communities. To propose that the community consisted solely of free neoplastic cells in lymph would be pure fancy. The community, in fact, would have been the lymph node. The damning aspect would have been an occult nodal metastasis, and not some fanciful stasis of neoplastic cells in lymph near the biopsy site. The subsequent behavior of such a melanoma would have been predicated on the intrinsic properties of melanoma cells, not only beyond the boundaries of the initial excision, but also within the primary lesion prior to the initial biopsy.
Fanciful virtual images are embodied in the concept of "in transit metastases." The concept give recognition to multiple dermal, and subcutaneous (not deep), metastases along the course of lymphatic drainage between the primary excision site and the regional lymph nodes. Distinctions between "in transit metastases" and local satellitosis have been difficult to define. Those based solely on the distance between the primary excision site, and the cutaneous metastases (12) have proved to be indiscriminate in practice.
The concept of "in transit" metastasis evolved as an explanation for the appearance of cutaneous metastases in limited distribution away from the excision site of a primary melanoma. A disease-free interval between the dissection of histologically negative lymph nodes, and subsequent appearance of local cutaneous metastases was basic to the concept. The concept of "in transit" metastases provided the rationale for a brief delay between the excision of a primary melanoma and a prophylactic, or therapeutic, lymph node dissection. The delay was designed to allow time for any cells in the lymphatic circulation to find their way to regional lymph nodes. Ideally, the delay would reduce the likelihood of subsequent regional cutaneous metastases. It presumably would allow "free-floating" (virtual) cells, with a potential for local implantation, time to find their way to the nodes prior to the node dissection. Hopefully, a subsequent node dissection, and any lymphatic obstruction related to the procedure, would not then lead to the implantation of cells locally in the pattern of satellitosis. If this concept had validity, there would be more than a little evidence favoring the utility of delayed prophylactic node dissections. The concept of "in transit" metastasis antedates the modern era in which histologic parameters dictate therapeutic decisions. Unfortunately, the concept, both muddled and ill- defined, has survived. The concept of "in transit" metastasis with its implications of malignant cells free in the lymphatic circulation, and uninfluenced, for considerable intervals, by the direction of the flow of lymph has little utility in the conceptualization of the biology of melanoma, and has no well-documented base.
Virtual images are indispensable in the interpretation of histologic sections, and in the practice of surgical pathology. Embroiled virtual and real images, elicited both by preconditioning, and by the examination of histologic sections, are to be reconciled, and incorporated, in the final interpretations. That the final reconciliations will provide an accurate portrayal of the nature of disease, or a reliable prediction of clinical course is always problematical.
The practice of pathology is a life-long endeavor. In it, the stores of virtual images require constant renewal. Of the stores, those, that are rarely used, tend to dim; they need to be refreshed at intervals. On the other hand, some virtual images, having once been evoked, are hard to lose, or dismiss, even when detrimental.
In this examination of the interplay of images, pathologists, clinicians, and patients, a new "standard of care" has not been defined. For a singular endeavor such as surgical pathology, "standards of care" must be individually defined. They are no more, nor less, than the excellence each pathologist demands of himself, and brings to his daily practice; they are as variable as the quality of practice provided by each of us.
Herein, the observations and admonitions are as individual as the histologic techniques to which they pertain. The drama of disease, and how the practices of the medical profession impact on disease, are both natural phenomena. The imposition of man (as a physician - an instrument) in the evolution of natural phenomena does not negate the naturalness of his interventions. As natural phenomena, the results of man's efforts are not always predictable; they need not to even be desirable, if measured by our appreciation of nature's apparent needs. Man's efforts, particularly those based on cognition, rather than intuition, may seem to be a perversion of natural phenomena but, if man is truly native here, his perturbations of nature, however destructive, are also natural. Even though nature is beyond blame, the legal industry would hold man, in his role as an instrument of the medical profession and, in turn, of nature, accountable for the vicissitudes of natural phenomena.
If pathology is an art, where are the masterpieces? The materials of pathology are ignoble. The real and virtual images of pathologists are representations of disease. Emotions, expressed as fear; concern; the insecurity of ill-health; the desperation that comes with the prospect of dying; and even the horror of death, all color the virtual images of patients. Pathologists' interpretations may only exacerbate the emotional state of patients, and further color their virtual images. The materials available to, and produced by, pathologists are not the stuff of masterpieces.
Dogged tenacity is required both in the manipulation of virtual images, and in the selection of relative linguistic symbols. In the efforts, even real images may become vicissitudinous; what once may have seemed appropriate may later seem absurd. In the practice of surgical pathology, virtual images, and respective labels shadow humility.
In the borderland of the neoplastic phenomena (as manifested in the dysplasia- melanoma sequence), our rhetoric fails to delineate sharply defined divisions. In all borderline categories, too few neoplasms behave in a malignant fashion to validate the designation, malignancy. In fact, the true definition of the potential of an individual example of all neoplasms is nature's secret, to be revealed at her leisure, and with no concern for man's manipulations of real and virtual images, or statistics. Time is the final arbiter.
Virtual images of other pathologists and even those of clinicians are routinely imposed upon consultants. These extraneous images, as embodied in the reports which accompany slides, often influence the consultants' approaches to problem cases. They may enrich those of the consultants but, on occasion, may muddle the interpretative process. In either case, prejudices are inherent in consultations. The consultant must evaluate not only real images, but also must deal with the virtual images, as embodied in the comments, and questions of submitters. For the most part, the extraneous images are an aid in the evocation of relevant virtual images.
The rhetoric of pathology is influenced by fads. As a consequence, our obser- vations, and our descriptions and designations tend to become stereotyped - they become repetitious. The very act of accommodating our rhetoric imposes restrictions on the evocation of virtual images. We tend to categorize diseases at a glance at low magnifications, and then must work to find other pertinent, or conflicting, real and virtual images. It is humiliating to discover that a precipitous selection of virtual images has been proved to be inappropriate and that, in turn, the respective rhetoric as documented in a report does not relate to the actual nature of the disease process.
Our degree of accuracy, and expertise, aside, biologic phenomena are poorly served by any form of rhetoric. Neither pathologists in their reports; experts in their scenarios; molecular pathologists in their techniques; nor attorneys in their expositions, really touch on nature's intent. Perhaps, a trifling with the hopes and frustrations of patients is the most damnable aspect of the rhetoric of modern medicine. Diagnosis and treatment currently are the main aspirations of the new generation of practitioners. Emotional responses to the needs of patients and their families are practically nonexistent. In managed care, they are inappropriate; the menial position of physicians in managed care casts emotional concerns as an extravagance.
The richness, and the utility, of individual stores of virtual images reflect the attention that has been paid to real images in the daily practice of pathology. The more detailed the histologic examination, the greater the fund of real images with which to construct, and then store, new virtual images. In turn, virtual images are our guides in the practice of pathology, and in the service of our patients, and fellow- physicians.
Expertise in histologic interpretations and in the conceptualization of disease processes is predicated on emotional contributions. In the examination of a histologic section, an "intuitive feeling” for the nature of the disease is based on the interplay of observations, knowledge, and emotions. A certainty of the appropriateness of a histologic diagnosis is a contribution of the emotions, not the intellect. Emotions, and the utility of virtual images, are integrally related. Emotions greatly influence the utility of pattern analysis in formulating a diagnosis. Often, after an examination utilizing the technique of pattern analysis and, at the level of a scanning lens, it is possible to characterize an initial histologic impression as a "feeling” for a particular category, or diagnosis. The romance between pathologist, and histologic preparations has been, and is, nurtured by virtual images, and their emotional overlay. In managed care, this romance will likely fade.
On occasion, a pathologist may have difficulty eliciting an emotional response to histologic real images. As a consequence, he may intellectualize the histologic interpretation. In this process, he mechanically evaluates the sum of the real images and, with no emotional overlay, constructs a diagnosis. He comes away from the task with no "feeling” that the histologic interpretation is appropriate. He may return to the same material several time over a period of days in search for the confidence instilled by an emotional concurrence. Cautions in the final categorization of such interpretations must be observed. They are to be relayed to the clinician, either verbally, or as a written report, or the histologic material from the problem lesion should be shared with other pathologists, at least one of whom, hopefully, will find the emotional store to confidently characterize the lesion. To ignore contributions by emotions would move us farther from an understanding of our interpretative processes and, in turn, of a patient's disease and of nature. Such a prospect is not inviting.
In histologic interpretations of difficult lesions, another borderland, namely that between the intellect and emotion, is exposed. For some pathologists, the pros- pect, that interpretations might be colored by emotions is disturbing. For them, the results of histologic interpretations should be purely a product of the intellect. These pathologists currently are promoting sophisticated checklists as a alternative to our emotion-rich, interpretative processes. Such checklists are being promoted as a way to avoid the capriciousness of the common interpretative process, and to standardize histologic interpretations. Utilizing such check lists, relative values could easily be assigned to the numbered items, and the values then factored to formulate numerical results. The numerical results, as an assigned numerical designation, might be translatable as “guides” to therapy. The greatest benefit from assigned numerical designations would be the rote nature of the process, one that is devoid of emotional overlays. The checklist approach might even lend itself to instrumentation with the human factor totally removed from the diagnostic endeavor. It then would lend itself to socialized medicine in which the needs of all must be met with a uniformity that would preclude the utilization, or even the availability, of all of the options in the practice of modern medicine (13). Exceptional care would cease to be an option. With such restricted, and controlled, systems, it then would be inappropriate to speak of an art of pathology.
1. Rambo ON: The limitations of histologic diagnosis. In, Progress in Radiation Therapy, vol II, edited by Buske F, Grune & Stratton, New York, 1962, pp. 215-24.
2. Hanson NR: Patterns of Discovery. Cambridge Press, New York, 1958.
3.Reed RJ: The histological variance of malignant melanoma: the interrelationship of histologic subtype, neoplastic progression, and biologic behavior. Pathol 1985; 17:301-12.
4. Ackerman AB: Macular and patch lesions of malignant melanoma: malignant melanoma in situ. J DermatolSurgOncol 1983; 9:615-9. 5. Clark WH Jr: From L, Bernadino EA, Mihm MC Jr: The histogenesis and biologic behavior of primary human malignant melanoma of the skin. Cancer Res 1969; 29:705-26.
6. Reed RJ: Minimal deviation melanoma. In, Pathobiology and Recognition of Malignant Melanoma, Chapter 9, edited by Mihm MC JR, Murphy GF, and Kaufman N, Williams &Wilkins, Baltimore, 1988, pp 110-52.
7. Reed RJ: Minimal deviation melanoma. Hum Pathol 1990; 21:1206-11.
8. Breslow A: Thickness, cross-sectional area and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970; 902-70.
9. Ronan SG, Eng AM, Briele HA, Shioura NN, Das Gupta: Thin malignant melanoma with regression and metastases. Arch Dermatol 1987; 123:1326-30.
10. Kornberg R, Harris M, Ackerman AB: Epidermotropically metastatic malignant melanoma: differentiating malignant melanoma metastatic to the epidermis from malignant melanoma primary in the epidermis. Arch Dermatol 1978; 114:67-9.
11. Smith JL, Stehlin: Spontaneous regression of primary malignant melanoma with metastases. Cancer 1965; 18:1399-1415. 12. Singletary SE, Balch CM: Recurrent regional metastases and their management. In Balch CM, et al, Cutaneous Melanoma, 2nd ed., JB Lippincott Co., Philadelphia, 1992, pp 428-9.
13. Dworkin R: will Clinton's plan be fair? The New York Review of Books.1993; 41:20-5.
14. Blessing K, Evans AT, Al-Nafussi A: Verrucousnaevoid and keratotic melanoma: a clinico-pathological study of 20 cases. Histopathol 1993; 23:453-8.
15. McNutt NS, Urmacher C, Hakimian J, Hoss DM, Lugo J: Nevoid malignant melanoma: morphologic patterns and immunohistochemical reactivity. J CutanPathol 1995; 22:502-17.
16. Reed RJ, Ichinose H, Clark WH Jr, Mihm MC Jr: Common and uncommon melanocytic nevi and borderline melanoma. SeminOncol 1975; 2:119-47.
17. Reed RJ: Minimal deviation melanoma arising in a congential nevus. Am J SurgPathol 1978; 2:215-20.
18. Reed RJ, Webb SV, Clark WH Jr: Minimal deviation melanoma (halo nevus variant). Am J SurgPathol 1990; 21:1206-11.
19. Reed RJ: Giant congenital nevua: a conceptualization of patterns. J Invest Dermatol 1993; 100:300s-12s.
20. Smith KJ, Barrett TL, Skelton HG, et al: spindle cell and epithelioid cell nevi with atypia and metastasis (malignant Spitz nevus). Am J SurgPathol 1989; 13:931-9.
21. Wong TY, Suster S, Duncan LM, Mihm MC Jr: Nevoid melanoma.
Hum Pathol 1995; 26:171-9.
22. Garcia Caballero T, Fraga M, Antunez JR, Perez Bacerra E,
Beiras A, Forteza J: Myxoid metastatic melanoma. Histopathol
23. Prieto VG, Kanik A, Salob S, McNutt NS: Primary cutaneous
myxoid malignant melanoma: immunohistologic clues to a difficult diagnosis. J Am AcadDermatol 1994; 30:335-9.
24. Bhuta S, Mirra JM, Cochran AJ: Myxoid malignant melanoma: a previously undiscribedhistologic pattern noted in metastatic lesions and a report of four cases. Am J SurgPathol 1986;10:203-11.